Texas Instruments TMS320C6000 Programmer's Manual page 183

Software Pipelining
Example 6–28. Assembly Code for Fixed-Point Dot Product (Software Pipelined
With No Extraneous Loads)
LDW .D1
|| LDW .D2
|| MVK .S1
|| ZERO .L1
|| ZERO .L2
[A1] SUB .S1
|| LDW .D1
|| LDW .D2
[A1] SUB .S1
||[A1] B .S2
|| LDW .D1
|| LDW .D2
[A1] SUB .S1
||[A1] B .S2
|| LDW .D1
|| LDW .D2
[A1] SUB .S1
||[A1] B .S2
|| LDW .D1
|| LDW .D2
MPY .M1X
|| MPYH .M2X
||[A1] SUB .S1
||[A1] B .S2
|| LDW .D1
|| LDW .D2
MPY .M1X
|| MPYH .M2X
||[A1] SUB .S1
||[A1] B .S2
|| LDW .D1
|| LDW .D2
6-46
ADDSPs all execute exactly 50 times. (The shaded areas of Example 6–29 in-
dicate the changes in this code.)
Executing the dot product code in Example 6–29 with no extraneous LDDWs
still requires a total of 74 cycles (9 + 41 + 9 + 15), but the code size is now larg-
er.
*A4++,A2 ; load ai & ai+1 from memory
*B4++,B2 ; load bi & bi+1 from memory
43,A1 ; set up loop counter
A7 ; zero out sum0 accumulator
B7 ; zero out sum1 accumulator
A1,1,A1 ; decrement loop counter
*A4++,A2 ;* load ai & ai+1 from memory
*B4++,B2 ;* load bi & bi+1 from memory
A1,1,A1 ;* decrement loop counter
LOOP ; branch to loop
*A4++,A2 ;** load ai & ai+1 from memory
*B4++,B2 ;** load bi & bi+1 from memory
A1,1,A1 ;** decrement loop counter
LOOP ;* branch to loop
*A4++,A2 ;*** load ai & ai+1 from memory
*B4++,B2 ;*** load bi & bi+1 from memory
A1,1,A1 ;*** decrement loop counter
LOOP ;** branch to loop
*A4++,A2 ;**** load ai & ai+1 from memory
*B4++,B2 ;**** load bi & bi+1 from memory
A2,B2,A6 ; ai * bi
A2,B2,B6 ; ai+1 * bi+1
A1,1,A1 ;**** decrement loop counter
LOOP ;*** branch to loop
*A4++,A2 ;***** ld ai & ai+1 from memory
*B4++,B2 ;***** ld bi & bi+1 from memory
A2,B2,A6 ;* ai * bi
A2,B2,B6 ;* ai+1 * bi+1
A1,1,A1 ;***** decrement loop counter
LOOP ;**** branch to loop
*A4++,A2 ;****** ld ai & ai+1 from memory
*B4++,B2 ;****** ld bi & bi+1 from memory